
import sys


import re
_r = re.escape
def _re_replace(s : str, to_replace : dict):
    for p, r in to_replace.items():
        s = re.compile(p).sub(r, s)
    return s


import matplotlib as mpl
import matplotlib.pyplot as plt
%matplotlib inline
%config InlineBackend.figure_formats = ['svg']
mpl.rcParams['figure.dpi'] = 300


!{sys.executable} -m pip install --quiet pandas
import pandas as pd


!{sys.executable} -m pip install --quiet seaborn
import seaborn as sns


!{sys.executable} -m pip install --quiet statsmodels numpy
import statsmodels.api as sm
import statsmodels.formula.api as smf
import statsmodels.tools as smt

!{sys.executable} -m pip install --quiet numpy
import numpy as np


!{sys.executable} -m pip install --quiet sklearn
import sklearn as skl

!{sys.executable} -m pip install --quiet scikit-learn-intelex
import sklearnex as sklx
#sklx.patch_sklearn()

import sklearn.linear_model
import sklearn.tree
import sklearn.ensemble 
import sklearn.model_selection 

from sklearn.compose import ColumnTransformer
from sklearn.pipeline import Pipeline
from sklearn.preprocessing import StandardScaler, OneHotEncoder

from sklearn.linear_model import LinearRegression
from sklearn.linear_model import Lasso

from sklearn.model_selection import train_test_split, GridSearchCV
from sklearn.metrics import mean_squared_error


from sklearn.model_selection import learning_curve

def plot_learning_curve(
    estimator,
    title,
    X,
    y,
    axes=None,
    ylim=None,
    cv=None,
    n_jobs=None,
    train_sizes=np.linspace(0.1, 1.0, 5),
):
    """
    Generate 3 plots: the test and training learning curve, the training
    samples vs fit times curve, the fit times vs score curve.

    Parameters
    ----------
    estimator : estimator instance
        An estimator instance implementing `fit` and `predict` methods which
        will be cloned for each validation.

    title : str
        Title for the chart.

    X : array-like of shape (n_samples, n_features)
        Training vector, where ``n_samples`` is the number of samples and
        ``n_features`` is the number of features.

    y : array-like of shape (n_samples) or (n_samples, n_features)
        Target relative to ``X`` for classification or regression;
        None for unsupervised learning.

    axes : array-like of shape (3,), default=None
        Axes to use for plotting the curves.

    ylim : tuple of shape (2,), default=None
        Defines minimum and maximum y-values plotted, e.g. (ymin, ymax).

    cv : int, cross-validation generator or an iterable, default=None
        Determines the cross-validation splitting strategy.
        Possible inputs for cv are:

          - None, to use the default 5-fold cross-validation,
          - integer, to specify the number of folds.
          - :term:`CV splitter`,
          - An iterable yielding (train, test) splits as arrays of indices.

        For integer/None inputs, if ``y`` is binary or multiclass,
        :class:`StratifiedKFold` used. If the estimator is not a classifier
        or if ``y`` is neither binary nor multiclass, :class:`KFold` is used.

        Refer :ref:`User Guide <cross_validation>` for the various
        cross-validators that can be used here.

    n_jobs : int or None, default=None
        Number of jobs to run in parallel.
        ``None`` means 1 unless in a :obj:`joblib.parallel_backend` context.
        ``-1`` means using all processors. See :term:`Glossary <n_jobs>`
        for more details.

    train_sizes : array-like of shape (n_ticks,)
        Relative or absolute numbers of training examples that will be used to
        generate the learning curve. If the ``dtype`` is float, it is regarded
        as a fraction of the maximum size of the training set (that is
        determined by the selected validation method), i.e. it has to be within
        (0, 1]. Otherwise it is interpreted as absolute sizes of the training
        sets. Note that for classification the number of samples usually have
        to be big enough to contain at least one sample from each class.
        (default: np.linspace(0.1, 1.0, 5))
    """
    if axes is None:
        _, axes = plt.subplots(1, 3, figsize=(20, 5))

    axes[0].set_title(title)
    if ylim is not None:
        axes[0].set_ylim(*ylim)
    axes[0].set_xlabel("Training examples")
    axes[0].set_ylabel("Score")

    train_sizes, train_scores, test_scores, fit_times, _ = learning_curve(
        estimator,
        X,
        y,
        cv=cv,
        n_jobs=n_jobs,
        train_sizes=train_sizes,
        return_times=True,
    )
    train_scores_mean = np.mean(train_scores, axis=1)
    train_scores_std = np.std(train_scores, axis=1)
    test_scores_mean = np.mean(test_scores, axis=1)
    test_scores_std = np.std(test_scores, axis=1)
    fit_times_mean = np.mean(fit_times, axis=1)
    fit_times_std = np.std(fit_times, axis=1)

    # Plot learning curve
    axes[0].grid()
    axes[0].fill_between(
        train_sizes,
        train_scores_mean - train_scores_std,
        train_scores_mean + train_scores_std,
        alpha=0.1,
        color="r",
    )
    axes[0].fill_between(
        train_sizes,
        test_scores_mean - test_scores_std,
        test_scores_mean + test_scores_std,
        alpha=0.1,
        color="g",
    )
    axes[0].plot(
        train_sizes, train_scores_mean, "o-", color="r", label="Training score"
    )
    axes[0].plot(
        train_sizes, test_scores_mean, "o-", color="g", label="Cross-validation score"
    )
    axes[0].legend(loc="best")

    # Plot n_samples vs fit_times
    axes[1].grid()
    axes[1].plot(train_sizes, fit_times_mean, "o-")
    axes[1].fill_between(
        train_sizes,
        fit_times_mean - fit_times_std,
        fit_times_mean + fit_times_std,
        alpha=0.1,
    )
    axes[1].set_xlabel("Training examples")
    axes[1].set_ylabel("fit_times")
    axes[1].set_title("Scalability of the model")

    # Plot fit_time vs score
    fit_time_argsort = fit_times_mean.argsort()
    fit_time_sorted = fit_times_mean[fit_time_argsort]
    test_scores_mean_sorted = test_scores_mean[fit_time_argsort]
    test_scores_std_sorted = test_scores_std[fit_time_argsort]
    axes[2].grid()
    axes[2].plot(fit_time_sorted, test_scores_mean_sorted, "o-")
    axes[2].fill_between(
        fit_time_sorted,
        test_scores_mean_sorted - test_scores_std_sorted,
        test_scores_mean_sorted + test_scores_std_sorted,
        alpha=0.1,
    )
    axes[2].set_xlabel("fit_times")
    axes[2].set_ylabel("Score")
    axes[2].set_title("Performance of the model")

    return plt


_data_ = {
    name: pd.read_csv(
        file, 
        engine = 'c',
        low_memory = True,
        memory_map = False, # set `False` to load into memory
        **kwargs
    ) for name, file, kwargs in [
        ('appearances', 'data/appearances.csv', {
            'dtype': {
                'player_id': 'object',
                'game_id': 'object',
                'appearance_id': 'object',
                'competition_id': 'object',
                'player_club_id': 'object'
            }
        }),
        ('clubs', 'data/clubs.csv', {
            'dtype': {
                'club_id': 'object'
            }
        }),
        #('competitions', 'data/competitions.csv', {}),
        ('games', 'data/games.csv', {
            'dtype': {
                'game_id': 'object'
            }
        }),
        #('leagues', 'data/leagues.csv', {}),
        ('players', 'data/players.csv', {
            'parse_dates': ['date_of_birth'],
            'dtype': {
                'player_id': 'object',
                'country_of_birth': 'category',
                'country_of_citizenship': 'category',
                'position': 'category',
                'sub_position': 'category'
            }
        }),
        ('player_valuations', 'data/player_valuations.csv', {
            'parse_dates': ['date'],
            'dtype': {
                'player_id': 'object'
            }
        })
    ]
}


data = {}


# clubs
data['clubs'] = _data_['clubs'].copy()

data['clubs'] = data['clubs'][[
    'club_id', 
    'pretty_name'
]]
data['clubs'].rename(
    columns = {'pretty_name': 'club_name'},
    inplace = True
)
data['clubs'].set_index('club_id', inplace = True)

data['clubs']


# games
data['games'] = _data_['games'].copy()

data['games'] = data['games'][[
    'season', 
    'game_id'
]]
data['games'].set_index('game_id', inplace = True)

data['games']


# appearances
data['appearances'] = _data_['appearances'].copy()

data['appearances'] = data['appearances'].loc[
    :, ~data['appearances'].columns.isin([
        'appearance_id', 
        'competition_id'
    ])
]
data['appearances'].rename(
    columns = {'player_club_id': 'club_id'},
    inplace = True
)

data['appearances'] = (
    data['appearances']
        .merge(
            data['games'], 
            on = 'game_id',
            copy = False
        ).drop(columns = 'game_id')
        .merge(
            data['clubs'], 
            on = 'club_id',
            copy = False
        ).drop(columns = 'club_id')
)

data['appearances'] = (
    data['appearances']
        .groupby(['player_id', 'season'])
        .agg({
            **{
                c: 'sum' for c in [
                    'goals', 
                    'assists', 
                    'minutes_played', 
                    'yellow_cards', 
                    'red_cards'
                ]
            },
            'club_name': 'last'
        })
        .reset_index('season')
)

data['appearances']


# player valuations
data['player_valuations'] = _data_['player_valuations'].copy()

data['player_valuations']['season'] = (
    pd.DatetimeIndex(data['player_valuations']['date']).year
)
data['player_valuations'].drop(columns = 'date', inplace = True)

data['player_valuations'] = (
    data['player_valuations']
        .groupby(['player_id', 'season'])
        .agg({'market_value': 'mean'})
        .reset_index('season')
)
data['player_valuations'].rename(
    columns = {'market_value_in_gbp': 'market_value'},
    inplace = True
)

data['player_valuations']


# players
data['players'] = _data_['players'].copy()

data['players'] = data['players'].loc[
    :, ~data['players'].columns.isin([
        'last_season',
        'name',
        'current_club_id',
        'market_value_in_gbp',
        'highest_market_value_in_gbp',
        'country_of_birth',
        'url', 
        'foot'
    ])
]
data['players'].rename(
    columns = {
        'pretty_name': 'name',
        'height_in_cm': 'height',
        'country_of_citizenship': 'nationality'
    },
    inplace = True
)

data['players']['sub_position'] = (
    data['players']['sub_position'].cat
        .rename_categories(
            lambda s: (
                _re_replace(s, {
                    fr'''^(.*){_r(' - ')}(.*)$''': r'\2'
                })
                .title()
            )
        )
)

data['players'].set_index('player_id', inplace = True)

data['players']


# final dataset
data['all'] = data['players'].merge(
    data['player_valuations'].merge(
        data['appearances'], 
        on = ['player_id', 'season'],
        copy = False
    ), 
    on = 'player_id',
    copy = False
)

data['all']['age'] = (
    pd.to_datetime(data['all']['season'], format = '%Y', utc = True) 
        - pd.to_datetime(data['all']['date_of_birth'], utc = True)
).astype('timedelta64[Y]')
data['all'].drop(columns = 'date_of_birth', inplace = True)

data['all'].dropna(axis = 'index', inplace = True)
data['all'].drop(
    data['all'][
        (data['all']['market_value'] <= 0)
            | (data['all']['height'] <= 0)
            | (data['all']['age'] <= 0)
    ].index, 
    inplace = True
)

data['all']


data['all'][data['all'].isna().any(axis = 1)]


data['all'].dtypes

name                object
nationality       category
position          category
sub_position      category
height               int64
season               int64
market_value       float64
goals                int64
assists              int64
minutes_played       int64
yellow_cards         int64
red_cards            int64
club_name           object
age                float64
dtype: object


data['all'].describe()


pd.DataFrame(data['all']['sub_position'].unique())


data['all'][data['all']['name'] == 'Cristiano Ronaldo']


data['all_eda'] = data['all'].copy()

data['all_eda']['log_market_value'] = np.log(data['all_eda']['market_value'])


df_highest_market_value_players = data['all_eda'].nlargest(n = 1, columns = 'market_value')

df_highest_market_value_players


df_highest_market_value = data['all_eda'].loc[data['all_eda']['name'].isin(df_highest_market_value_players['name'])]

df_highest_market_value


_ = sns.barplot(
    data = df_highest_market_value, 
    x = 'season', y = 'market_value'
).set(
    title = f'''{df_highest_market_value_players['name'].values}'s '''
    '''Market Value Over the Years'''
)
plt.xticks(rotation = 45, ha = 'right', rotation_mode = 'anchor')

plt.show()


_, axs = plt.subplots(ncols = 2)

_ = sns.histplot(
    ax = axs[0],
    data = data['all_eda'].reset_index(), 
    x = 'market_value', 
    bins = 50
).set(title = 'Distribution of Market Values')

_ = sns.histplot(
    ax = axs[1],
    data = data['all_eda'].reset_index(), 
    x = 'log_market_value', 
    bins = 50
).set(title = 'Distribution of Log Market Values')

plt.tight_layout()
plt.show()


_ = sns.boxplot(data = data['all_eda'][[
    'goals',
    'assists',
    'yellow_cards',
    'red_cards'
]]).set(title = 'Distribution of Goals, Assists, Yellow/Red Cards')


_ = sns.histplot(
    data = data['all_eda'].reset_index(), x = 'height',
    bins = 50
).set(title = 'Distribution of Height')


corr = data['all_eda'].corr()
_ = sns.heatmap(corr,  
    cmap = 'seismic', 
    linewidth = 1, linecolor = 'white',
    vmax = 1, vmin = -1,
    mask = np.triu(np.ones_like(corr, dtype = bool)), 
    annot = True,
    fmt = '0.2f'
)


_ = sns.pairplot(
    data['all_eda'].sample(frac = .05).reset_index(), 
    plot_kws = {'rasterized': True}
)


model_ols = smf.ols(formula = '''
    market_value ~ 
        age + goals + assists + minutes_played 
            + yellow_cards + red_cards 
            + height + age 
''', data = data['all_eda'])

fit = model_ols.fit()
pred = fit.predict()

fit.summary()


df_pred = data['all_eda'].assign(market_value_pred = pred)

df_pred


print('RMSE', smt.eval_measures.rmse(
    df_pred['market_value'], 
    df_pred['market_value_pred']
))

RMSE 7334580.4334724955


model_ols = smf.ols(formula = '''
    log_market_value ~ 
        age + goals + assists + minutes_played 
            + yellow_cards + red_cards 
            + height + age 
''', data = data['all_eda'])

fit = model_ols.fit()
pred = fit.predict()

fit.summary()


df_pred = data['all_eda'].assign(log_market_value_pred = pred)

df_pred[['log_market_value', 'log_market_value_pred']]


print('RMSE', smt.eval_measures.rmse(
    df_pred['log_market_value'], 
    df_pred['log_market_value_pred']
))

RMSE 1.283259436824852


model_ols = smf.ols(formula = '''
    log_market_value ~ 
        age + goals + assists + minutes_played 
            + yellow_cards + red_cards 
            + height + age 
            + C(nationality) + C(position) + C(sub_position) + C(club_name)
''', data = data['all_eda'])

fit = model_ols.fit()
pred = fit.predict()

fit.summary()


df_pred = data['all_eda'].assign(market_value_pred = np.exp(pred), log_market_value_pred = pred)

df_pred


print('RMSE', smt.eval_measures.rmse(
    df_pred['log_market_value'], 
    df_pred['log_market_value_pred']
))

RMSE 0.8584421552905203


def diagnostic_plot(X, y):    
    rgr = LinearRegression()
    rgr.fit(X, y)
    pred = rgr.predict(X)

    plt.subplot(1, 3, 1)
    plt.scatter(pred, y, alpha=0.1)
    plt.plot(y, y, color = 'red', linewidth = 1)
    plt.title('Regression fit')
    plt.xlabel('Predicted y')
    plt.ylabel('y')

sns.set(style = 'darkgrid')
diagnostic_plot(
    X = data['all_eda'][[
        'goals', 
        'assists', 
        'minutes_played', 
        'yellow_cards', 
        'red_cards', 
        'height',
        'age'
       # 'nationality',
       # 'position',
       # 'sub_position',
       # 'club_name'
    ]], 
    y = data['all_eda']['log_market_value']
)


model_ols = smf.ols(formula = '''
    log_market_value ~ 
        C(nationality) + C(position) + C(sub_position) + C(club_name)
''', data = data['all_eda'])

fit = model_ols.fit()
pred = fit.predict()


print('RMSE', smt.eval_measures.rmse(
    data['all_eda']['log_market_value'], 
    pred
))

RMSE 0.9917601841201406


# feature selection
X = data['all_eda'][[
    'goals', 
    'assists', 
    'minutes_played', 
    'yellow_cards', 
    'red_cards', 
    'height',
    'age',
    'season',
    'nationality',
    'position',
    'sub_position',
    'club_name'
]]
y = data['all_eda']['log_market_value']

# train-test split
X_train, X_test, y_train, y_test = \
    train_test_split(X, y, test_size = 0.2, random_state = 0)

preprocessor = ColumnTransformer(
    transformers = [
        ("num", StandardScaler(), [
            'goals',
            'assists',
            'minutes_played', 
            'yellow_cards', 
            'red_cards', 
            'height',
            'age'
        ]),
        ("cat", OneHotEncoder(handle_unknown = 'ignore'), [
            'nationality', 
            'position', 
            'sub_position', 
            'club_name'
        ])
    ]
)


model_lasso = Pipeline(
    steps = [
        ('preprocessor', preprocessor), 
        ('dtr', skl.linear_model.Lasso())
    ]
)


model_lasso.fit(X_train, y_train)
model_lasso_y_pred = model_lasso.predict(X_test)

print('accuracy', model_lasso.score(X_test, y_test))
print('MSE', mean_squared_error(y_test, model_lasso_y_pred))

accuracy -2.901700328017398e-05
MSE 2.254265463732437


model_dtr = Pipeline(
    steps = [
        ('preprocessor', preprocessor), 
        ('dtr', skl.tree.DecisionTreeRegressor())
    ]
)


model_dtr.fit(X_train, y_train)
model_dtr_y_pred = model_dtr.predict(X_test)

print('accuracy', model_dtr.score(X_test, y_test))
print('MSE', mean_squared_error(y_test, model_dtr_y_pred))

accuracy 0.2822816498513936
MSE 1.6178807433761904


model_cv_dtr = skl.model_selection.GridSearchCV(
    model_dtr, 
    param_grid = {
        'dtr__splitter': ['best', 'random'], 
        #'dtr__max_depth': [1, 3, 5, 7, 9, 11, 12], 
        'dtr__max_depth': [1, 5, 9, 12], 
        #'dtr__min_samples_leaf': [1, 2, 3, 4, 5, 6, 7, 8, 9, 10], 
        'dtr__min_weight_fraction_leaf': [0.1, 0.2, 0.3, 0.4, 0.5], 
        'dtr__max_features': ['auto', 'log2', 'sqrt', None], 
        #'dtr__max_leaf_nodes': [None, 10, 20, 30, 40, 50, 60, 70, 80, 90]
    },
    scoring = 'neg_mean_squared_error',
    cv = 5,
    #verbose = 3,
    n_jobs = -1
)


model_cv_dtr.fit(X_train, y_train).score(X_test, y_test)

-1.6765851975675852


model_cv_dtr.best_params_

{'dtr__max_depth': 5,
 'dtr__max_features': 'auto',
 'dtr__min_weight_fraction_leaf': 0.1,
 'dtr__splitter': 'best'}


model_rfr = Pipeline(
    steps=[('preprocessor', preprocessor), 
           ('rfr', skl.ensemble.RandomForestRegressor(max_depth = 5))]
)


model_rfr.fit(X_train, y_train).score(X_test, y_test)

0.33956166004816013


model_cv_rfr = skl.model_selection.GridSearchCV(
    model_rfr, 
    param_grid = {
        'rfr__n_estimators': [10, 50, 100],
        'rfr__max_depth': [1, 5, 9, 12], 
        'rfr__min_weight_fraction_leaf': [0.1, 0.3, 0.5], 
        'rfr__max_features': ['auto', 'log2', 'sqrt', None]
    },
    scoring = 'neg_mean_squared_error',
    cv = 5,
    #verbose = 3,
    n_jobs = -1
)


model_cv_rfr.fit(X_train, y_train).score(X_test, y_test)

-1.6521475613018508


model_cv_rfr.best_params_

{'rfr__max_depth': 12,
 'rfr__max_features': 'auto',
 'rfr__min_weight_fraction_leaf': 0.1,
 'rfr__n_estimators': 100}


plot_learning_curve(
    model_cv_rfr, 
    title = 'Learning Curve (Random Forest)', 
    X = X, y = y, 
    n_jobs = -1
)

	club_name
club_id
1032	Fc Reading
2323	Orduspor
1387	Acn Siena 1904
3592	Kryvbas Kryvyi Rig
1071	Wigan Athletic
...	...
1269	Pec Zwolle
200	Fc Utrecht
317	Fc Twente Enschede
3948	Royale Union Saint Gilloise
1304	Heracles Almelo

	season
game_id
2244388	2012
2219794	2011
2244389	2012
2271112	2012
2229332	2012
...	...
3646190	2021
3646188	2021
3655616	2021
3655629	2021
3646191	2021

	season	goals	assists	minutes_played	yellow_cards	red_cards	club_name
player_id
10	2014	32	18	4578	12	0	Lazio Rom
10	2015	16	14	3428	6	0	Lazio Rom
100009	2014	0	0	5576	8	0	Kuban Krasnodar
100009	2015	2	2	4512	12	0	Kuban Krasnodar
100009	2016	0	0	1260	6	0	Anzhi Makhachkala
...	...	...	...	...	...	...	...
99923	2014	0	2	832	4	0	Cagliari Calcio
99924	2016	0	2	1824	6	0	Ca Osasuna
99977	2014	0	0	194	0	0	Rcd Mallorca
99977	2015	10	6	3046	2	0	Royal Excel Mouscron
99977	2019	0	0	716	0	0	Caykur Rizespor

	season	market_value
player_id
10	2004	6300000.0
10	2005	10800000.0
10	2006	22500000.0
10	2007	20700000.0
10	2008	18000000.0
...	...	...
99977	2018	990000.0
99977	2019	720000.0
99977	2020	562500.0
99977	2021	495000.0
99977	2022	540000.0

	name	nationality	date_of_birth	position	sub_position	height
player_id
254016	Arthur Delalande	France	1992-05-18	Midfield	Central Midfield	186
51053	Daniel Davari	Iran	1988-01-06	Goalkeeper	Goalkeeper	192
31451	Torsten Oehrl	Germany	1986-01-07	Attack	Centre-Forward	192
44622	Vladimir Kisenkov	Russia	1981-10-08	Defender	Right-Back	182
30802	Oscar Diaz	Spain	1984-04-24	Attack	Centre-Forward	183
...	...	...	...	...	...	...
462285	Fabian De Keijzer	Netherlands	2000-05-10	Goalkeeper	Goalkeeper	193
368612	Merveille Bokadi	DR Congo	1996-05-21	Defender	Centre-Back	186
408574	Joey Veerman	Netherlands	1998-11-19	Midfield	Central Midfield	185
364245	Jordan Teze	Netherlands	1999-09-30	Defender	Centre-Back	183
575367	Richard Ledezma	United States	2000-09-06	Attack	Attacking Midfield	174

COGS 118A - Final Project¶

Group members¶

Abstract¶

Background¶

Problem Statement¶

Data¶

Evaluation¶

Postprocessing and Exploratory Data Analysis¶

Market Value Distribution (Single player)¶

Market Value Distribution (Overall)¶

Proposed Solution¶

Evaluation Metrics¶

Results¶

Subsection 1¶

Subsection 2¶

Subsection 3¶

Categorical + Numerical¶

Categorical ONLY¶

Subsection 4¶

Lasso Regression¶

Decision Tree + Grid Search¶

Random Forest + Grid Search¶

Discussion¶

Limitations¶

Ethics & Privacy¶

Conclusion¶

Footnotes¶

	name	nationality	position	sub_position	height	season	market_value	goals	assists	minutes_played	yellow_cards	red_cards	club_name	age
player_id
9800	Artem Milevskyi	Ukraine	Attack	Centre-Forward	189	2020	90000.0	0	0	720	6	0	Fk Minaj	34.0
43084	Gaetano Berardi	Switzerland	Defender	Right-Back	179	2020	360000.0	0	0	228	0	0	Leeds United	31.0
230826	Gennaro Acampora	Italy	Midfield	Central Midfield	174	2020	360000.0	2	4	1248	4	0	Spezia Calcio	25.0
198087	Matteo Ricci	Italy	Midfield	Defensive Midfield	176	2020	1530000.0	0	6	4880	10	0	Spezia Calcio	25.0
110689	Deniz Mehmet	Turkey	Goalkeeper	Goalkeeper	192	2020	68000.0	0	0	1080	0	0	Dundee United Fc	27.0
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
364245	Jordan Teze	Netherlands	Defender	Centre-Back	183	2019	420000.0	0	0	360	0	0	Psv Eindhoven	19.0
364245	Jordan Teze	Netherlands	Defender	Centre-Back	183	2020	1102500.0	0	2	7494	10	0	Psv Eindhoven	20.0
364245	Jordan Teze	Netherlands	Defender	Centre-Back	183	2021	5400000.0	2	8	5260	12	0	Psv Eindhoven	21.0
575367	Richard Ledezma	United States	Attack	Attacking Midfield	174	2020	658250.0	0	2	234	2	0	Psv Eindhoven	19.0
575367	Richard Ledezma	United States	Attack	Attacking Midfield	174	2021	765000.0	2	0	88	0	0	Psv Eindhoven	20.0

	height	season	market_value	goals	assists	minutes_played	yellow_cards	red_cards	age
count	50365.000000	50365.000000	5.036500e+04	50365.000000	50365.000000	50365.000000	50365.000000	50365.000000	50365.000000
mean	182.287938	2017.381813	3.654199e+06	3.897151	2.964400	2816.958801	6.026645	0.143989	24.658791
std	6.616766	2.315522	8.302298e+06	7.367342	4.805991	2102.487008	6.100280	0.545081	4.357635
min	159.000000	2013.000000	9.000000e+03	0.000000	0.000000	2.000000	0.000000	0.000000	14.000000
25%	178.000000	2015.000000	3.600000e+05	0.000000	0.000000	900.000000	2.000000	0.000000	21.000000
50%	183.000000	2017.000000	9.000000e+05	2.000000	2.000000	2580.000000	4.000000	0.000000	24.000000
75%	187.000000	2019.000000	3.150000e+06	4.000000	4.000000	4420.000000	10.000000	0.000000	28.000000
max	206.000000	2021.000000	1.800000e+08	122.000000	62.000000	10122.000000	46.000000	6.000000	42.000000

	name	nationality	position	sub_position	height	season	market_value	goals	assists	minutes_played	yellow_cards	red_cards	club_name	age
player_id
8198	Cristiano Ronaldo	Portugal	Attack	Centre-Forward	187	2014	96000000.0	122	46	9282	12	2	Real Madrid	28.0
8198	Cristiano Ronaldo	Portugal	Attack	Centre-Forward	187	2015	105000000.0	102	30	8586	6	0	Real Madrid	29.0
8198	Cristiano Ronaldo	Portugal	Attack	Centre-Forward	187	2016	99000000.0	84	24	8252	10	0	Real Madrid	30.0
8198	Cristiano Ronaldo	Portugal	Attack	Centre-Forward	187	2017	90000000.0	88	16	7356	10	0	Real Madrid	31.0
8198	Cristiano Ronaldo	Portugal	Attack	Centre-Forward	187	2018	96000000.0	56	20	7292	8	2	Juventus Turin	32.0
8198	Cristiano Ronaldo	Portugal	Attack	Centre-Forward	187	2019	74250000.0	70	14	7982	6	0	Juventus Turin	33.0
8198	Cristiano Ronaldo	Portugal	Attack	Centre-Forward	187	2020	54000000.0	76	8	7682	10	0	Juventus Turin	34.0
8198	Cristiano Ronaldo	Portugal	Attack	Centre-Forward	187	2021	39000000.0	48	6	6202	20	0	Juventus Turin	35.0

	name	nationality	position	sub_position	height	season	market_value	goals	assists	minutes_played	yellow_cards	red_cards	club_name	age	log_market_value
player_id
342229	Kylian Mbappe	France	Attack	Centre-Forward	178	2019	180000000.0	48	24	4102	4	0	Fc Paris Saint Germain	20.0	19.008467

Dep. Variable:	market_value	R-squared:	0.220
Model:	OLS	Adj. R-squared:	0.219
Method:	Least Squares	F-statistic:	2023.
Date:	Fri, 10 Jun 2022	Prob (F-statistic):	0.00
Time:	18:50:46	Log-Likelihood:	-8.6764e+05
No. Observations:	50365	AIC:	1.735e+06
Df Residuals:	50357	BIC:	1.735e+06
Df Model:	7
Covariance Type:	nonrobust

	coef	std err	t	P>\|t\|	[0.025	0.975]
Intercept	-3.263e+06	9.5e+05	-3.434	0.001	-5.13e+06	-1.4e+06
age	-7.257e+04	7704.492	-9.419	0.000	-8.77e+04	-5.75e+04
goals	2.387e+05	5594.966	42.664	0.000	2.28e+05	2.5e+05
assists	3.224e+05	9256.579	34.825	0.000	3.04e+05	3.41e+05
minutes_played	816.2531	23.146	35.265	0.000	770.886	861.620
yellow_cards	-6.42e+04	6966.312	-9.216	0.000	-7.79e+04	-5.05e+04
red_cards	-8.313e+04	6.08e+04	-1.368	0.171	-2.02e+05	3.6e+04
height	2.699e+04	5171.108	5.219	0.000	1.69e+04	3.71e+04

Omnibus:	52944.353	Durbin-Watson:	0.637
Prob(Omnibus):	0.000	Jarque-Bera (JB):	5493249.031
Skew:	5.172	Prob(JB):	0.00
Kurtosis:	53.106	Cond. No.	1.02e+05

Dep. Variable:	log_market_value	R-squared:	0.277
Model:	OLS	Adj. R-squared:	0.276
Method:	Least Squares	F-statistic:	2750.
Date:	Fri, 10 Jun 2022	Prob (F-statistic):	0.00
Time:	18:50:46	Log-Likelihood:	-84026.
No. Observations:	50365	AIC:	1.681e+05
Df Residuals:	50357	BIC:	1.681e+05
Df Model:	7
Covariance Type:	nonrobust

Omnibus:	829.490	Durbin-Watson:	0.849
Prob(Omnibus):	0.000	Jarque-Bera (JB):	780.323
Skew:	0.267	Prob(JB):	3.59e-170
Kurtosis:	2.707	Cond. No.	1.02e+05

	log_market_value	log_market_value_pred
player_id
9800	11.407565	13.285919
43084	12.793859	13.119897
230826	12.793859	13.524196
198087	14.240778	14.502799
110689	11.127263	13.274930
...	...	...
364245	12.948010	12.977332
364245	13.913091	14.922000
364245	15.501910	14.675605
575367	13.397340	13.038863
575367	13.547631	12.973216

	coef	std err	t	P>\|t\|	[0.025	0.975]
Intercept	12.6996	0.166	76.389	0.000	12.374	13.025
age	0.0146	0.001	10.794	0.000	0.012	0.017
goals	0.0235	0.001	23.997	0.000	0.022	0.025
assists	0.0447	0.002	27.589	0.000	0.042	0.048
minutes_played	0.0003	4.05e-06	63.687	0.000	0.000	0.000
yellow_cards	8.021e-05	0.001	0.066	0.948	-0.002	0.002
red_cards	0.0104	0.011	0.979	0.328	-0.010	0.031
height	-0.0005	0.001	-0.553	0.580	-0.002	0.001

	coef	std err	t	P>\|t\|	[0.025	0.975]
Intercept	11.5308	0.571	20.178	0.000	10.411	12.651
C(nationality)[T.Albania]	0.0734	0.622	0.118	0.906	-1.146	1.292
C(nationality)[T.Algeria]	0.2551	0.621	0.411	0.681	-0.962	1.472
C(nationality)[T.Angola]	0.3074	0.624	0.492	0.622	-0.916	1.531
C(nationality)[T.Antigua and Barbuda]	0.2160	0.714	0.302	0.762	-1.183	1.615
C(nationality)[T.Argentina]	0.2977	0.619	0.481	0.631	-0.916	1.512
C(nationality)[T.Armenia]	-0.0014	0.625	-0.002	0.998	-1.227	1.224
C(nationality)[T.Aruba]	-0.0902	0.652	-0.138	0.890	-1.368	1.188
C(nationality)[T.Australia]	0.1675	0.622	0.269	0.788	-1.052	1.387
C(nationality)[T.Austria]	0.2851	0.621	0.459	0.646	-0.931	1.502
C(nationality)[T.Azerbaijan]	-0.1142	0.643	-0.178	0.859	-1.374	1.146
C(nationality)[T.Bahrain]	-1.712e-13	1.79e-14	-9.562	0.000	-2.06e-13	-1.36e-13
C(nationality)[T.Barbados]	-2.56e-13	1.83e-13	-1.397	0.163	-6.15e-13	1.03e-13
C(nationality)[T.Belarus]	0.0723	0.624	0.116	0.908	-1.151	1.296
C(nationality)[T.Belgium]	0.1595	0.619	0.258	0.797	-1.054	1.373
C(nationality)[T.Benin]	-0.0629	0.633	-0.099	0.921	-1.304	1.178
C(nationality)[T.Bermuda]	0.8274	1.065	0.777	0.437	-1.259	2.914
C(nationality)[T.Bolivia]	0.2710	0.665	0.408	0.683	-1.032	1.574
C(nationality)[T.Bosnia-Herzegovina]	0.1339	0.621	0.216	0.829	-1.083	1.351
C(nationality)[T.Brazil]	0.2422	0.619	0.391	0.696	-0.971	1.456
C(nationality)[T.Bulgaria]	0.2262	0.626	0.362	0.718	-1.000	1.452
C(nationality)[T.Burkina Faso]	0.2372	0.625	0.379	0.704	-0.988	1.462
C(nationality)[T.Burundi]	0.2987	0.656	0.455	0.649	-0.987	1.584
C(nationality)[T.Cameroon]	0.2567	0.620	0.414	0.679	-0.959	1.473
C(nationality)[T.Canada]	-0.0501	0.626	-0.080	0.936	-1.277	1.177
C(nationality)[T.Cape Verde]	0.1182	0.623	0.190	0.849	-1.103	1.339
C(nationality)[T.Central African Republic]	0.4542	0.647	0.702	0.483	-0.814	1.723
C(nationality)[T.Chad]	0.1665	0.665	0.251	0.802	-1.136	1.469
C(nationality)[T.Chile]	0.2355	0.623	0.378	0.705	-0.986	1.457
C(nationality)[T.China]	-0.4146	0.656	-0.632	0.527	-1.700	0.871
C(nationality)[T.Chinese Taipei (Taiwan)]	-0.7022	0.761	-0.923	0.356	-2.193	0.789
C(nationality)[T.Colombia]	0.4535	0.621	0.731	0.465	-0.763	1.670
C(nationality)[T.Comoros]	0.3605	0.640	0.563	0.573	-0.894	1.616
C(nationality)[T.Congo]	0.0272	0.627	0.043	0.965	-1.201	1.256
C(nationality)[T.Costa Rica]	0.1420	0.626	0.227	0.821	-1.085	1.369
C(nationality)[T.Cote d'Ivoire]	0.3630	0.620	0.586	0.558	-0.852	1.578
C(nationality)[T.Croatia]	0.3909	0.620	0.631	0.528	-0.824	1.606
C(nationality)[T.Cuba]	-0.0319	1.066	-0.030	0.976	-2.122	2.058
C(nationality)[T.Curacao]	-0.0264	0.626	-0.042	0.966	-1.253	1.200

Omnibus:	2544.681	Durbin-Watson:	1.353
Prob(Omnibus):	0.000	Jarque-Bera (JB):	3861.538
Skew:	-0.450	Prob(JB):	0.00
Kurtosis:	4.016	Cond. No.	2.83e+19